I'm interested in income and social mobility. I understand that Piketty's research deals with the the distribution of income over a long time period. I have also seen some of Chetty and Saez's research concerning income and social mobility in the US. I'm wondering, what data do Chetty and Saez work use? What other data is potentially available and how far back does it go?

It would be nice to see who could find the data that goes back the farthest. (Also, the data does not necessarily have to be US data.)

  • $\begingroup$ The PSID is standard, I'm guessing you want something longer? $\endgroup$
    – jayk
    Nov 21, 2014 at 2:01
  • 1
    $\begingroup$ I emailed the team behind the linked article, let's see if I get a response. :) $\endgroup$ Nov 21, 2014 at 17:15
  • $\begingroup$ Both graphs come from federal tax return data. This has to come through the IRS, and in the past I've been told that you can get this data "but only if you basically sign over your soul." Understandably people have privacy concerns about the IRS handing over their tax returns to researchers en masse. $\endgroup$
    – jayk
    Nov 21, 2014 at 21:36
  • $\begingroup$ Also, I'm pretty sure you need security clearance and hence US citizenship. $\endgroup$
    – FooBar
    Nov 22, 2014 at 18:12
  • 1
    $\begingroup$ On this subject, obs.rc.fas.harvard.edu/chetty/NSFdataaccess.pdf $\endgroup$
    – jayk
    Nov 24, 2014 at 17:12

2 Answers 2


Maybe check Clark's work. He allegedly uses long time series. http://jasoncollins.org/2014/09/30/the-genetic-basis-of-social-mobility/


Regarding income inequality:

Economists used to rely on household surveys in which individuals (or households) reported their earnings, social benefits, etc. Although this is still done, recent studies rely more and more on the income data recorded in the tax files due to its larger coverage and accuracy (especially at the top of the distribution).

You can find the most complete historic series of income (& wealth) inequality based on tax records in the World Wealth and Income Inequality Database: http://wid.world/fr/accueil/

Other sources to compare income inquality and poverty levels across countries and time are: - The World Bank: https://data.worldbank.org/ - The UNU-WIDER World Income Inequality Database: https://www.wider.unu.edu/project/wiid-world-income-inequality-database - The Standardized World Income Inequality Database http://fsolt.org/swiid/ - Branko Milanovic's data of Gini coefficients in 1950-2022 for 170 countries: http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EXTRESEARCH/0,,contentMDK:22301380~pagePK:64214825~piPK:64214943~theSitePK:469382,00.html

For nice visualizations over time you can see Gap Minder: - https://www.youtube.com/watch?v=jbkSRLYSojo

Regarding intergenerational income mobility (~social mobility):

The first studies of intergenerational income mobility (i.e. how income levels compare across several generations) relied on household surveys in which retrospective information about parents' income was collected. Then, panel surveys started to be used when they had a sufficient coverage (panel surveys follow the same individuals -and often their children- over time).

Some examples of long-panels available are: - PSID for the US (1968-Today): https://psidonline.isr.umich.edu/ - BCS for the UK (1970-Today): https://bcs70.info/ - BHPS for the UK (1991-Today): https://www.iser.essex.ac.uk/bhps - SOEP for Germany (1984-Today): https://www.diw.de/en/soep

In the last decades some studies of intergenerational mobility have also used income data coming from tax records, linking the files of parents' and their children (see the pioneering work of Miles Corak (1999) for Canada). That is the kind of data that Chetty and coauthors have used.

Gregory Clark and coauthors have been relying on rare surnames and their over-representation in top institutions (e.g. Cambridge and Oxford) to track social mobility over centuries. You can check this work in his website.

Just mentioning that there are works looking at intergenerational mobility of other dimensions besides income, such as wealth (e.g. Clark and Cummins, 2015), educational attainment (e.g. Lindhal et al., 2015), or even attitudes (Dohmen et al., 2011).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.